Universality in the bulk holds close to given points

From MaRDI portal
Publication:2275486





The universality limit for the distribution of eigenvalues of random \(n \times n\) Hermitian matrices is studied. A new method to establish the universality is presented based on classical complex analysis such as normal families and the theory of entire functions of exponential type. Let \(\mu\) be a measure with compact support and a point \(\xi\) lies in that support while \(\mu'\) is positive and continuous at \(\xi\). Let \(A_n\) be a sequence of positive numbers with \(\lim_{n \to \infty} {A_n} =~\infty\). Then it is shown that it is possible to choose \(\xi_n \in [\xi - \frac{A_n}{n},\xi + \frac{A_n}{n}]\) in such a way that \[ \lim_{n \to \infty} \frac {K_n \left( \xi_n, \xi_n + \frac{a}{\tilde K_n(\xi_n, \xi_n)} \right)} {K_n(\xi_n, \xi_n)} = \frac{\sin \pi a}{\pi a} \] uniformly for a compact subset of the plane. In the given formula \(K_n\) means the \({n}\)th reproducing kernel for \(\mu\) and \(\tilde K_n\) is its normalized kernel. This universality in bulks holds for varying points close to a given point \(\xi\) without assuming regularity for \(\mu\). Similar results are valid also for sequences of measures.



Cites work









This page was built for publication: Universality in the bulk holds close to given points

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2275486)